Esr Study of Solutions of Sulfur and Polysulfides in Liquid Ammonia V

ESR STUDY OF SOLUTIONS OF SULFUR AND POLYSULFIDES IN LIQUID AMMONIA V. Pinon, E. Levillain, A. Demortier, J. Lelieur

To cite this version:

V. Pinon, E. Levillain, A. Demortier, J. Lelieur. ESR STUDY OF SOLUTIONS OF SULFUR AND POLYSULFIDES IN LIQUID AMMONIA. Journal de Physique IV Proceedings, EDP Sciences, 1991, 01 (C5), pp.C5-223-C5-230. 10.1051/jp4:1991526. jpa-00250650

HAL Id: jpa-00250650 https://hal.archives-ouvertes.fr/jpa-00250650 Submitted on 1 Jan 1991

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. JOURNAL DE PHYSIQUE IV Colloque C5, supplement au Journal de Physique I, Vol. 1, decembre 1991 C 5 - 2 2 3

ESR STUDY OF SOLUTIONS OF SULFUR AND POLYSULFIDES IN LIQUID AMMONIA

V. PINON, E. LEVILLAIN, A. DEMORTIER and J.P. LELIEUR URA 253 CNRS, HEI, 13 rue de Toul, F-59046 Lille cedex, France

Résumé : L'étude RPE des solutions de soufre et de polysulfures dans l'ammoniac liquide montre la présence d'une seule raie lorentzienne, située à g = 2.0292. Ce signal doit être attribué au radical anion S3" identifié dans ces solutions par spectrophotométrie et par spectroscopie Raman, et qui est donc la seule espèce radicalaire dans ces solutions. La concentration du radical S3" a été déduite des expériences RPE pour une large gamme de concentration des solutions. Pour les différentes solutions étudiées, la concentration de S3" augmente avec la concentration de la solution et passe par un maximum observé pour des solutions concentrées et attribué qualitativement à un appartement des spins des radicaux S3".

Abstract : An ESR study of solutions of sulfur and polysulfides in liquid ammonia gives evidence of a simple Lorentzian line located at g = 2.0292. This signal must be assigned to the radical anion S3" identified in these solutions by spectrophotometry and Raman spectroscopy. It is shown that S3" is the only radical species in these solutions. The concentration of S3" has been determined from the ESR experiments for a wide concentration range of the solutions. For the various solutions, the concentration of S3" increases with the concentration of the solution and goes through a maximum, located in the concentrated range of solutions. This maximum is qualitatively interpreted by a spin pairing mechanism between the S3" radicals.

Introduction A significant step in the understanding of the solutions of sulfur in liquid ammonia (SAS) was the identification in these solutions of S4N" and of the radical anion S3" by Chivers and Lau III using Raman spectroscopy. However an ESR signal had not been previously observed in these solutions 111. Chivers and Lau attributed the lack of an ESR signal in SAS to a possible dimerization of the S3" radical. The identification of S3" in SAS was confirmed by Bernard et al. /3,4/ using Raman spectroscopy and 2 spectrophotometry. Dubois et al. 75,6/ showed that S6 ~ is the least reduced polysulfide in liquid

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jp4:1991526 C5-224 JOURNAL DE PHYSIQUE IV ammonia, and it is partly dissociated into S3-. The dissociation of S62- is observed only for temperatures higher than ca. 200 K, and is strongly temperature dependent. Dubois et al. /5,6/ also observed that polysulfides are disproportionated in liquid ammonia, and that the disproportionation is higher for acidic solutions. In neutral solutions (Li2S,), the S3- radical has been observed from Raman spectroscopy and from spectrophotometry for n > 3. It is not observed for Li2S3-NH3 solutions (for instance) because S32- is not disproportionated in neutral solutions. However, in acidic solutions ,(NH4)2Sn ,the S3- radical has been observed for n > 1. Such differences result from the different disproportionation in neutral and acidic solutions. The purpose of the present paper is to report the ESR identification of S3- in the various ammonia solutions in which this radical has been previously observed with other techniques : sulfur, lithium polysulfides and ammonium polysulfides. The purpose of these ESR experiments was also to check if S3- is the only radical in the investigated solutions, and also to determine the concentration of S3- in rather concentrated solutions, because for these highly colored solutions the concentration of S3- can only be deduced from the absorption spectra for dilute solutions.

Experimental The lithium and ammonium polysulfide solutions are prepared by reducing sulfur with lithium and hydrogen suEde respectively. The preparation of the solutions /5,6/ , the ESR cells and the experimental conditions 1131 have been previously described.

9 ,8.. Figure 1. ESR signal of S3- at 290 K in a Li2S6 - NH3 solution (4 10-3 M). The experimental signal is 0 0 0 0 parabolicfitted by baseline.the sum of a Lorentzian lineshape and a :< I 9~ .: . . . . :experimental data points. o o o : Lorentzian lineshape :parabolic baseline t~.~~~~.~.14.4~~~~8ul~~u~~~*.~t~~~8~ 1500 2500 9500 4500 The sensitivity on the vertical scale is four times MAGNETIC FIELD (GAUSS) larger for the residuals than for the ESR signal.

The experimental signal is always rather weak and this leads to use high receiver gains. In these conditions a drift of the baseline is observed. The parameters of the ESR lineshape have been determined by fitting the experimental signal (Fig. 1) to the sum of a parabolic curve describing the drift of the baseline and of a Lorentzian signal following 171 : where H, is the magnetic field at the center of the line, AHpp the peak-to-peak linewidth of the derivative, Y the maximun of the derivative ;a, b and c are the parameters describing the parabolic shape of the baseline. The fit of f(H) to the experimental signal using a non-linear least squares technique allows the determination of H,, AHpp, Y, a, b and c. These parameters allow the calculation of the area A, of the absorption signal using the classical formula I71 relative to a Lorentzian lineshape :

The determination of the area A, was found more accurate with the use of this fit procedure and Eq.(2) than with the numerical double integration technique available from the Bruker ESP 1600 computer program. The numerical double integration technique leads systematically to lower values for A, by ca.5 %. This is assigned to the broad linewidth of the Lorentzian lineshape. On the wings of the signal, the signal is merged in the noise over a large magnetic field range.

Figure 2. Plot of A, versus A. The area A, is given for an internal diameter of 1 mm and the absorbance A for an optical pathlength of 0.05 cm.

ABSORBANCE

The conversion of the ESR signal of a given sample into the concentration of the paramagnetic species requires the calibration of the sensitivity of the spectrometer. For this purpose, we have demonstrated that the ESR signal of the Sg- radical anion observed in solution of lithium hexasulfde (Li2S6) in liquid ammonia can be used as a standard 181. It can easily be shown I81 that the area A, of the absorption signal is related to the absorbance A at 610 nm by : C5-226 JOURNAL DE PHYSIQUE IV where E is the extinction coefficient of S3- ,1 the optical path length, V the volume of sample undergoing the ESR resonance phenomenon, and P is a proportionality constant Consequently, if for samples at a given temperature, or for a sample at various temperature, A, is plotted versus the absorbance A at 610 nm, a linear variation must be obtained. Such a linear variation is found as shown in Fig. 2.

Results. General asDea For all the investigated samples (sulfur and polysulfides in ammonia) a single ESR line has been observed for temperatures higher than 200 K. All the recorded lines can be fitted to a Lorentzian lineshape. The position of the line (g = 2.0292 +. 0.0009) is temperature and concentration independent. This ESR signal must be assigned to S3- because this radical has previously been identified in these solutions with Raman spectroscopy and visible spectrophotometry. For all the investigated samples, the area A, of the ESR signal, i.e. the concentration of S3-, decreases when the temperature decreases and the ESR signal is not detectable at 200 K (Figs. 3a and 3b). These observations are in agreement with the variations of the concentration of S3- with temperature observed by spectrophotometry in dilute solutions of sulfur or polysulfides in liquid ammonia.

Figure 3. ESR signal of S3- ;A : in a S - NH3 solution (0.1 M) ;T= 286,255 , 225 K. ;B : in a Li2S6 solution (4 10-3 M) ; T = 290,280,270, ... 200 K.

Solutions of sulfur in ammonia, The ESR signal has been detected in S-NH3 in a large concentration range from about 1 10-3 M up to saturated solution (3.6 M) between 200 and 290 K. At a given temperature, the variation of the concentration of S3- with the concentration of the solution displays an unexpected behavior (Fig. 4). In the dilute range of concentration, the concentration of S3- increases slowly with the concentration of the solution : for a 2.5 10-3 M solution, the concentration of Sg- is 3 10-4 M at room temperature. When the concentration of the solution is increased by a factor of one hundred (0.25 M), the concentration of S3- is increased by a factor of four and is close to 1.2 10-3 M. These data confm that the disproportionation of sulfur in ammonia is a chemical equilibrium, as it has been previously shown 11 1,121. i.e. the disproportionation of sulfur is not complete. The unexpected aspect of these results is that the concentration of S3- goes through a maximum for concentrated solutions (ca. 1.5 M, i.e. for a mole ratio MI3/ S - 24) 1131. A similar behavior has been observed in solutions of polysulfides in ammonia, as indicated and discussed below.

-2.00

elz 0 1 V 1.50 -

I+ E-c $1.00 1 npre4 Variation of the S3- concentration versus E t t the concentrarion of the S - NH, solution at 290 K. &v 4 0.50 - v0 * t +

Solutions of ~olvsulfidesin ammonia., For the comparison of the results obtained for polysulfide solutions with various stoichiometries ( i.e. for various values of n), it was found useful to have a common concentration scale for all these solutions. This concentration scale is defined relatively to the number of sulfur atoms in the solution, whatever its oxidation state. On this scale, the "molar concentration" is given by C, = 36 n / R. The concentration of S3-, deduced from the ESR signal for the various solutions at 290 K, is plotted versus the concentration of the solution on Fig. 5, where the concentration of the solution is given on the C, scale. As expected from the absorption spectra /5,6/, the concentration of S3- is larger in Li2S6-NH3 solutions than in Li2S4-NH3 solutions and in @iH4)2S6-NH3 than in (NI-I&S4-NH3. The ESR results also confm that the disproportionation of polysulfides ions is higher in acidic (NH4+)solutions that in neutral solutions. For n = 6 , the concentration of S3- is higher in lithium than in ammonium solutions, because ~6~-is disproportionated in ammonium solutions, as it has been shown from visible spectrophotometry/5,6/. For n = 4 , ~4~-is more disproportionated in acidic solutions and this leads to higher concentration of SG2-and therefore of S3- : the concentration of S3- is therefore higher in (NHq)2S4-m3 solutions than in Li2S4-NH3 solutions. These results previously shown from visible spectrophotometry for dilute solutions are here extended for all the concentrations. C5-228 JOURNAL DE PHYSIQUE IV

Figure 5. Variation of the concentration of S3- at 290 K versus the concentration C, of the solution in the various studied solutions. The relative uncertainty on the concentration of S3' is lower than 10 %. O Li2S6 - NH3 O (NH4)2S6 - NH3 Li2S4 - NH3 * (NH4I2S4 - NH3

Figure 5 also shows that, for all types of solutions of polysulfide, the concentration of S3- increases with the concentration of the solution and goes through a maximun. This maximum is found for concentrated solutions. On the C, scale, this maximum is located at the same concentration of the solution ( ca 1 M). It is found that the concentration of S3- at the maximum depends upon the studied solutions : at the maximum, the concentration of S3- is about five times larger for Li2S6-NH3 solutions than for Li2S4-NH3 solutions. Obviously, this maximun cannot be expected from the equilibrium constant given by eq. (3). It can be easily shown I91 that the concentration [S3-] can be expressed as a function of the dissociation equilibrium constant K(T) of Ss2- and the concentration C, of ~6~-at 200 K :

From eq. (4) the concentration [S3-1 increases monotically with C,. Obviously, this equation is strictly valid only for dilute solutions when the activity coefficients can be ignored. Let us admit that eq (4) gives the theoretical concentration of S3-, i.e. [S3-Ith , while the ESR experiments give the experimental concentration [S3-Iexp. The difference between [S3-Ith and [S3-lexp can be rationalized by plotting Log( [S3-Iexp I [S3-Ith ) versus (cs)lJ2 (Fig. 6). A plateau is observed for dilute solutions, because the concentration of S3- is then correctly predicted by eq. (4). For moderately concentrated and concentrated solutions, a linear variation is observed versus (c,)ln. This plot shows that the chemical process leading to the maximum of the concentration of S3- starts from the dilute range of concentration, and it suggests that the variations of the concentration of S3- result from the same chemical mechanism in all the concentration range, except for highly dilute solutions. The observed correlation between Log( [S3-lexp/ IS3-]& ) and (Cs)l12 suggests that we are dealing with a mechanism controlled by ion-ion interactions. However, at the microscopic level, the decrease of the ratio [S3-Iexp1 [S3-Ith can be assigned to a spin pairing process between the S3- radicals. These radicals are submitted to two opposite forces : the long range coulombic repulsion force and the short range attractive magnetic force which is the driving force to the spin pairing process. When the concentration of the solution increases, the average distance between the S3- radicals decreases and the probability that two radicals could be paired increases. It is suggested that when two radicals are close enough to be paired, they form a transient species in which each S3- keeps its structure. The same comparison between [S3-lexp I [S3-la and (QIR has been made for other types of polysulfide solutions in ammonia : (N%)2S6, (N%)2S4, Li2S4. Similar variations to those observed for Li2S6-NH3 solutions have been found, indicating that we are dealing with the same physical effect in all types of solutions 1141.

Figure 6. Log ( [S3-leXpI [S3-Ith ) versus cS1I2. The numerical equation of the linear part of the plot is : - 0.89 CS1l2+ 0.1 14

Variation of the concentration of S3- with temDerature, It is well known that when the temperature decreases, the equilibrium between ss2-and s3-is shifted towards s~~-.The activation energy AH of this equilibrium has been determined for dilute solutions from spectrophotometry experiments and found to be equal to 49.1 -e 1.0 kJ.mo1-l. If eq. (4) is considered, when 16 Co >> K(T) (i.e. Co >> 3 10-4 M at room temperature), the concentration [S3-1 is approximatively given by :

with KO= K, exp( - AH IRT). The activation energy can, therefore, be determined from the ESR experiments by plotting Log is3-]versus 1 I T. For dilute Li2S6-NH3 solutions, AH is found equal to 49.0 1.0 kJ.mol-l. However, for solutions more concentrated than 0.1 M, AH is found equal to ca. 60 k~.mol-l.The same values of AH are obtained for (NH4)2S6-NH3 solutions : about 50 k~.mol-Ifor dilute solutions and 60 kJ.mo1-1 for concentrated solutions. The similarity of the results obtained for LizS6 and (NH~)~S~solutions is an indication that the disproportionation of S62- in the latter solutions is temperature independent.

rature, The experimental values of AHpp for the various investigated solutions give a well defined curve if the concentration scale C, is used (Fig. 7). For dilute and moderately concentrated solutions, the linewidth is practically concentration independent. It is observed that the linewidth decreases sharply when the concentration of S3- decreases, i.e. for C, > 1 M. This correlation suggests that both variations could have the same origin. C5-230 JOURNAL DE PHYSIQUE IV

160

h m Figure 7. Variation of the linewidth AJ?$,P at 290 K 120 (d M versus the concentration C, of the solution for all the V 80 investigated solutions. g: x 0 Li2S6 - NH3 (NH4)2S6-NH3 4 a W Li2S4 - NH3 * (NHq)2S4 - NH3

It has been found that the linewidth increases when the temperature increases : approximately 10 G. for every 10 degrees 1141. The variations of Log AHpp are linear versus 1 I T, with a slope equal to 5.6 + 0.7 kJ.mol-1. This slope is constant for dilute and moderately concentrated solutions. A close correlation is found between AHw and E T I q. where E is the dielectric constant and q the viscosity of pure liquid ammonia. The slope of Log (E T l q) is found equal to 5.3 kJ.mol-1. The relaxation mechanisms will not be discussed in the present paper but it must be emphasized that the relaxation time is short ( - 1 ns) and that we are therefore dealing with a very fast relaxation mechanism.

Acknowledgements : The authors thank Professor G. Lepoutre for stimulating discussions. They express their gratitude to the CNRS, the R6gion Nord-Pas de Calais and the fondation Norbert Segard for fellowships to V.P. and E.L..

References I11 Chivers, T., Lau, C., Inorg. Chem. 21 (1982) 453 1Z Kerouanton, A, Herlem, M., Thiebault, A, Anal. Lett. 6 (1973) 171 I31 Bernard, L., Thhse de Docteur-Ingenieur, no 310, Lie(1982) I41 Bernard, L., Lelieur, J.P., Lepoutre, G., Nouv. J. Chim. 9 (1985) 199 I51 Dubois, P., Lelieur, J.P. Lepoutre, G., Inorg. Chem. 27 (1988) 73 I61 Dubois, P., Lelieur, J.P. Lepoutre, G., Inorg. Chem. 27 (1988) 1883 I71 Poole, C.P., Electron Spin Resonance, Wiley-Interscience, New York (1983) I81 Pinon, V., Levillain, E., Lelieur, J.P.. The S,- radical as a standard for ESR experiments to appear in the Journal of Magnetic Resonance 191 Pinon, V., Lelieur, J.P., Inorg. Chem. 30 (1991) 2260 I101 Chivers, T., Nature 32 (1974) 252 11 11 Dubois, P., Thhse, Universite? de Lille, No 184 (1987) IlZ Dubois, P., Lelieur, J.P. Lepoutre, G., Inorg. Chem. 28 (1989) 195 I131 Pinon, V., Levillain, E., Lelieur, J.P., J. Phys. Chem. 95 (1991) 6462 I141 Pinon, V., Demortier, A., Lelieur, J.P., unpublished results